⚛️ quantum physics

Bang-bang protocol for nondispersive qubit readout

This paper proposes and analyzes a "bang-bang" qubit readout protocol that utilizes sudden quenches of the coupling constant in the nondispersive regime to achieve fast, high-fidelity, and quantum-non-demolition single-shot measurements with errors scaling as $1/N$ , effectively overcoming the speed limitations of conventional dispersive approaches while avoiding measurement-induced state transitions.

Original authors: Nina del Ser, Yinan Chen, Jacob Steiner, Gil Refael

Published 2026-08-26

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Nina del Ser, Yinan Chen, Jacob Steiner, Gil Refael

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

In the world of quantum computing, the ability to read the state of a tiny information carrier, known as a qubit, is as critical as the ability to store it. These qubits are the fundamental units of quantum machines, capable of holding complex information that classical computers cannot manage. However, extracting this information without destroying it is a formidable challenge. If the measurement process is too slow or clumsy, the delicate quantum state collapses or changes before the data can be recorded. For years, scientists have relied on a method called dispersive readout, which is gentle and reliable but inherently slow. It works by listening to a faint shift in the frequency of a nearby resonator, a device that vibrates like a tiny tuning fork. While this approach preserves the qubit's state, the signal is weak, and the process takes time, creating a bottleneck for faster, more powerful quantum computers.

A team of researchers at the California Institute of Technology has proposed a radically different approach to solve this speed problem. They describe a new protocol, which they call "bang-bang" readout, that operates in a regime where the qubit and the resonator interact much more strongly than before. Instead of the slow, careful listening of the traditional method, this new technique involves a sudden, sharp switch-on of the connection between the qubit and the resonator, followed by an equally sudden switch-off. The resonator is preloaded with a large number of photons, or particles of light, before the connection is made. When the switch is flipped, the interaction happens so quickly that the two possible states of the qubit cause the resonator to move in two distinct, easily distinguishable directions almost instantly. The researchers found that this method allows for a measurement that is not only incredibly fast but also preserves the integrity of the qubit's state, a property known as quantum non-demolition.

The core of this discovery lies in how the researchers managed the sudden interaction. In the past, scientists believed that turning on such a strong connection abruptly would likely disturb the qubit, scrambling its information. The team, however, demonstrated that if the qubit is prepared in a specific orientation before the switch is flipped, the sudden interaction does not destroy the information. Instead, the qubit evolves in a predictable way, remaining in a nearly pure state while the resonator's state separates clearly based on the qubit's initial condition. By analyzing the mathematical dynamics of this system, the researchers showed that the errors introduced by this sudden "quench" of the connection become vanishingly small as the number of photons in the resonator increases. Specifically, the error decreases in proportion to the inverse of the number of photons, meaning that with enough photons, the measurement becomes nearly perfect.

To test their theory, the researchers developed a detailed analytical model based on the Jaynes-Cummings model, a standard description of how light and matter interact. They used a mathematical technique called a saddle-point approximation to handle the complex calculations involved with large numbers of photons. Their simulations revealed that with a typical coupling strength of 100 megahertz and a detuning of 23 megahertz, the readout could be completed in about 7 nanoseconds. This is significantly faster than the 50 nanoseconds or more often required by current dispersive methods. Furthermore, they found that the measurement fidelity—the accuracy with which the qubit's state is identified—and the quantum non-demolition quality both exceed 99 percent when the effective number of photons is around 11 or higher. This high level of accuracy ensures that after the measurement, the qubit remains in a known state and can be reset and reused for further calculations without needing to cool it down or wait for it to relax naturally.

The researchers also explored a way to improve this process even further by adding a second, classical drive to the qubit, essentially pushing it with an external field. They discovered that by carefully matching the phase of this external push with the light already in the resonator, they could effectively increase the number of photons interacting with the qubit without physically adding more light. This "effective" increase in photons further reduced the measurement errors and improved the purity of the qubit's state after the readout. This finding suggests that the protocol is flexible and can be optimized using existing control hardware, such as magnetic flux drives, rather than requiring entirely new physical components.

While the theoretical framework is robust and the simulations are consistent, the authors acknowledge that real-world quantum devices are more complex than the idealized models they used. Their work assumes a simple two-level system and does not yet account for all the complications of multi-level qubits or the specific noise present in actual superconducting circuits. They note that future work will need to address how these sudden switches might interact with other energy levels in the qubit or with the environment. However, the current study provides a strong theoretical foundation, proving that a fast, non-destructive readout is possible without the slow, adiabatic ramps that have been the standard for so long. By replacing gradual adjustments with sudden, precise actions, the "bang-bang" protocol offers a promising path toward the high-speed, high-fidelity measurements needed for the next generation of quantum technology.

Technical Summary: Bang-bang Protocol for Nondispersive Qubit Readout

Problem Statement
High-fidelity, fast, and quantum-non-demolition (QND) readout is a fundamental bottleneck in superconducting quantum computing. Conventional readout operates in the dispersive regime, where the qubit-resonator coupling $g$ is weak compared to the detuning $\Delta$ . While QND, this approach limits the readout rate to $\sim g^2\sqrt{N}/\Delta \ll g$ , where $N$ is the photon number. To achieve faster readout rates approaching the bare coupling scale $g$ , protocols must operate in the nondispersive regime. However, traditional nondispersive protocols rely on adiabatic parameter sweeps to load and unload the resonator. These sweeps traverse ranges of photon occupation that can encounter multiphoton resonances between dressed eigenladders, leading to measurement-induced state transitions (MIST) and parasitic coupling to two-level systems. The central challenge is to realize a fast, nondispersive readout that avoids these resonances while maintaining high QNDness.

Methodology
The authors propose a "bang-bang" readout protocol that replaces adiabatic sweeps with sudden quenches of the qubit-resonator coupling constant $g$ . The protocol proceeds as follows:

Initialization: The resonator is preloaded with a large coherent state $|\alpha_0\rangle$ (photon number $N = |\alpha_0|^2$ ) while decoupled from the qubit.
Quench On: The coupling $g$ is suddenly switched on to a value $g \gg \Delta/(2\sqrt{N})$ .
Evolution: The system evolves under the Jaynes-Cummings (JC) Hamiltonian for a duration $t_r \approx 4g^{-1}$ , allowing pointer states to separate.
Quench Off: The coupling is suddenly switched off.
Measurement: The resonator state is measured projectively (modeled as a binary projection on the momentum quadrature).

To analyze this non-perturbative dynamics, the authors develop an analytical theory for the JC model in the regime of large photon numbers ( $N \gtrsim 1$ ). Their approach combines:

Saddle-point approximation: Expanding observables around the global maximum of the initial coherent state's Poisson distribution.
Systematic perturbation theory: Expanding correction terms in powers of $1/\sqrt{N}$ .
Comparison with the Rabi model: Using the Rabi model (classical drive) as a baseline to understand the leading-order dynamics and identify deviations caused by quantum back-action.

The theory is extended to include a second, classical drive on the qubit (e.g., a flux drive), which is shown to be equivalent to increasing the effective photon number $N_{\text{eff}}$ .

Key Contributions and Results

Sweet-Spot Basis and Purity:
The authors identify a specific "sweet-spot" basis $\{|+\rangle, |-\rangle\}$ for the qubit, distinct from the computational basis $\{|\uparrow\rangle, |\downarrow\rangle\}$ . If the qubit is initialized in one of these states, the qubit-resonator entanglement is minimized.
- The qubit purity at the readout time $t_r$ remains close to unity, with a deviation scaling as $1/N$ . Specifically, $1 - \text{Tr}[\hat{\rho}_q^2(t_r)] \approx 1/(4N)$ for large $N$ .
- The post-measurement state is a known, time-dependent superposition $|\pm(t)\rangle$ , rather than a computational eigenstate.
Pointer State Separation:
In the nondispersive regime, the pointer states (resonator coherent states conditioned on the qubit state) separate at the bare coupling rate $g$ , independent of $N$ . This contrasts with the dispersive regime where the rate is suppressed by $\Delta$ . The separation time is $t_r \approx 4g^{-1}$ , which is significantly faster than dispersive readout times (e.g., $<10$ ns vs. $>50$ ns for typical parameters).
Fidelity and QNDness:
The authors define measurement fidelity ( $F$ ) and QNDness based on the post-measurement density matrices.
- Scaling: Both the error in fidelity and the error in QNDness scale as $1/N_{\text{eff}}$ (or $1/N$ without the classical drive).
- Performance: For typical parameters ( $g/2\pi = 100$ MHz, $\Delta/2\pi = 23$ MHz), the protocol achieves $>99\%$ QNDness and high fidelity for effective photon numbers $N_{\text{eff}} \gtrsim 11$ .
- Time Dependence: Unlike dispersive readout where QNDness remains stable over a wide window, the nondispersive QNDness degrades over time (scaling with $t^2$ ) due to higher-order corrections. However, for sufficiently large $N_{\text{eff}}$ , a "safe readout" window exists where QNDness remains above 99%.
Role of Classical Driving:
The paper demonstrates that applying a classical drive to the qubit (in phase with the resonator drive) effectively increases the photon number to $N_{\text{eff}} = |\alpha_0 + s|^2$ . This allows for improved readout performance (higher purity, lower error) without necessarily increasing the physical photon population in the resonator, thereby mitigating potential heating or leakage issues.

Significance and Claims
The paper claims that the bang-bang protocol successfully realizes a fast, single-shot projective measurement in the nondispersive regime without the need for adiabatic ramps, thereby avoiding the resonances that cause MIST.

Speed: The readout time is determined by the coupling strength $g$ rather than the dispersive shift, offering a potential order-of-magnitude speedup.
QNDness: Contrary to the intuition that non-adiabatic coupling destroys QNDness, the authors show analytically that QNDness is preserved parametrically at large $N$ , with errors vanishing as $1/N$ .
Theoretical Framework: The work provides a systematic analytical expansion for the JC model beyond the leading-order Gea-Banacloche approximation, capturing subleading corrections necessary to quantify measurement errors and purity loss.

Limitations and Future Work
The authors explicitly state that their current work is limited to a two-level qubit approximation within the rotating-wave approximation. They do not yet account for:

Counter-rotating terms (CR terms) arising from capacitive coupling or flux drives.
Multi-level qubit effects (leakage to non-computational states).
Finite damping and environmental decoherence.
The full experimental implementation of avoiding resonances in realistic multilevel devices post-quench.

The paper concludes that a detailed study of these realistic complexities is reserved for future work.

Technical Summary: Bang-bang Protocol for Nondispersive Qubit Readout

More like this